Bio

I believe that there might be three things that would be revolutionary in the 21st century: 1) the movement from the United States dollar standard to the cryptocurrency standard, 2) the replication of level of human-like perception in machines (artificial intelligence), and 3) synthetic biological manufacturing and analysis. To align with those goals, I decided to 1) cofound a cryptocurrency derivatives exchange, and 2) research, design and implement advanced deep learning algorithms for government clients and institutions. 3) spend to learn and donate to synthetic biological causes. My current research interest is the credit-assignment problem (alternatives to backpropagation). Experience: - sound event detection research with Tandemlaunch, - research on sound and speech synthesis in Silicon Valley, - research on conditional image synthesis with Microsoft Research Montreal, - dialogue systems MILABot team, - LSTM regularization and their implementations in C/C++ Specialization: - writing deep neural network layers on embedded small devices optimized using EIGEN and SIMD. Frameworks: - Theano (contributor), - LibGPUArray (contributor), - Pytorch, - Chainer, - Tensorflow, - Oríon